Introduces a unified RL policy that jointly handles force and position control on quadrupedal and humanoid robots without force sensors, enabling position+force tracking, compliant behaviors, and force-aware imitation learning for contact-rich tasks.
